B2B startup metrics: revenue, retention, and gross margin

Executive overview

Most early-stage founders either track too little (launching blind) or too much (500-metric dashboards before they have real data). Neither works. Pick four or five key metrics, define them precisely, and track them consistently from day one.

For B2B startups, three numbers belong at the top of every investor update: revenue, burn rate, and runway. Beyond those, retention and gross margin are the metrics that determine whether a business can compound or is just filling a leaky bucket.

A business with flat retention curves and net dollar retention above 100% compounds automatically — one without is an endless, futile scramble.

Setting up metrics before launch

• Pick four or five key metrics — not 30 or 50. The number grows over time.
• Use the simplest analytics tool available; SQL queries on your own database are fine.
• Write down precise definitions and get the whole team to agree on them.
• Constant internal arguments about definitions destroy meeting productivity.
• Inconsistent definitions also make cross-company comparisons meaningless.
• Don't change definitions when the numbers disappoint — you're only fooling yourself.

Revenue: the core B2B metric

• Vanity metrics — page views, unique visitors, GMV, gross transaction value — look big but don't reflect the health of the business.
• Optimising for GMV can mask flat or declining revenue (e.g. signing bigger customers with large cashback deals).
• For B2B companies, revenue is almost always the right primary metric.
• Don't hide bad revenue numbers; putting a prominent zero in an investor update keeps everyone honest and focused.
• Alongside revenue, always include burn rate (monthly costs minus revenue) and runway (months until cash runs out).

Retention and the cohort layer cake

• Retention measures how many customers from a given signup month are still paying in subsequent months.
• Track cohorts (January cohort, February cohort, etc.) and stack them visually to see the shape of the business.
• High retention creates a layer cake: cohorts stay fat, revenue accumulates, and the business grows even without new sales.
• Low retention creates a leaky bucket: customers signed up in month one are gone by month three, and growth requires perpetually replacing them.
• It matters more that retention flattens at some point than that it flattens at a high level — 20% stable retention beats 80% that decays to zero.

Net dollar retention for B2B SaaS

• Net dollar retention (NDR) measures revenue from a cohort a year later versus at signup, accounting for churn and upsells.
• Example: 10 customers at $10k/month = $100k MRR. One year later, two churn (−$20k), three upsell to $20k (+$30k) → $110k = 110% NDR.
• NDR above 100% means cohorts grow over time; below 100% means they shrink.
• Early-stage B2B SaaS should target NDR of 125–150%+, for three reasons:
  • Initial pricing is usually too low.
  • Product improves and justifies higher prices.
  • Sales and upsell skills improve over time.
• Mature B2B SaaS: 110–120% NDR is healthy. Below 100% signals a product problem — fix retention before investing in more top-of-funnel.

Gross margin and the danger of negative unit economics

• Gross margin = revenue minus cost of goods sold (costs that vary per customer).
• For AI products, API costs to providers like OpenAI or Anthropic are a real COGS — free credits just hide the cost temporarily.
• Pure SaaS historically ran at ~95% gross margin; operationally intensive businesses (delivery, installation, services) may sit at 5–15%.
• Low gross margin means you need far more revenue to cover fixed costs and reach profitability.
• Negative gross margin (charging less than the marginal cost to serve) can be a deliberate blitz-scaling strategy, but requires a credible plan to flip unit economics.
• Scale only after fixing negative unit economics — not before.
• The 2010–2021 zero-interest-rate era made negative-margin scaling popular; most of those businesses failed when capital dried up.

Applying metrics without hiding behind them

• Don't split-test minor decisions (button colour) at small scale — you lack the data volume for significance.
• Split-test consequential decisions: pricing tiers, packaging, key product flows.
• Metrics don't replace talking to customers. Staying close to users is non-negotiable.
• Use metrics, customer conversations, and product intuition together — not any one alone.